Taylor Rule Follow-up: Core CPI Is Not Biased

Paul Krugman made a valuable contribution on measuring inflation in light of my post on the latest from John Taylor:

Or at least it seems to be a new rule — namely, pick whatever price index makes the point you want, even if it’s not at all the price index you would normally use … Um, the inflation rate for the “GDP price index”? That’s the GDP deflator, which the Fed very carefully does not use as a policy indicator. Why? Because it contains things like grain and oil prices, which fluctuate a lot, so that it’s an unstable measure that is highly unreliable as an indicator of underlying inflation. The Fed prefers the consumption deflator excluding food and energy.

Taylor has a heated reply but alas it is all heat and no light:

Rather than taking out food and energy price inflation I controlled for price volatility in that rule by averaging overall inflation over time. Simply taking out food and energy price inflation can lead to policy errors especially when such inflation lasts for more than a short time. And it is not only the overall GDP price level. The CPI inflation rate was also rising, not falling, during this period. In any case, the increase rather than a decrease in overall inflation was only one part of my assessment that this was not a slack period. I also discussed the unemployment rate—which got quite low (4.4%) rather than high as in slack periods—and the huge housing boom with high housing price inflation.

OK – inflation rose but only slightly. A 4.4% unemployment rate was certainly not low in comparison to what we witnessed in the late 1990’s. And it is odd that Dr. Taylor refused to acknowledge my point that the FED had already been increasing interest rates before the labor market got moderately over the Bush recession. As far as the housing boom – which President Bush used to brag about – a lot of the blame should go to unwise financial market deregulation – but I guess it would be political suicide for a Republican economist to acknowledge that. In my view, Krugman reader JCB had a more interesting – albeit invalid (as I will explain) – set of comments:

Does eliminating the most volatile categories of consumer expenditure such as food, energy, and home prices from the calculation of the year-on-year rate of inflation make it possible to ignore them as important elements of the long term standard of living? I mean, why tacitly assume that the most volatile prices will sum to zero in the long run? ... The cumulative divergence between all-consumer prices and "underlying inflation" increases between 2000-2013

JCB provided us with his evidence, which was a chart showing how core CPI rose by a cumulative amount of 32% over the 2000 to 2013 period whereas CPI (including food and energy) rose by a cumulative amount of 38% during this same period. This period has often been described as the great commodities boom as noted by Pedro Conceição and Heloisa Marone:

The trough, since when the 21st century boom started, took place in late 2001. In real terms (using the US CPI to deflate the nominal price series), the boom remains impressive, with indices more than doubling in real terms. However, real prices were still below the average prices of the 1970s and earlier decades (Figure 2).

If one compared core CPI to overall CPI for the period from 1979 to 2001, core CPI rose faster than the overall consumer price index. Over the entire period, both series have increased by about 228%. If JCB was trying to suggest a long-term bias in the use of core CPI, I don’t see it. Rather – I just see more volatility in the use of an index that includes food and energy prices.